I am trying to create an xsl sheet that returns part of a tree from xml.
I need all page nodes with stop=0 up untill the node with attribute stop=1

stripped xml sample:

<content>
  <page id='page1' stop='0'>
    <page id='page2' stop='0'/>
    <page id='page3' stop='0'>
      <page id='page4' stop='0'/>
    </page>
  </page>
  <page id='page5' stop='0'>
    <page id='page6' stop='1'/>
    <page id='page7' stop='0'>
      <page id='page8' stop='1'/>
    </page>
  </page>
</content>

In this case I would like to have returned all pages from 1 to 6 (where it is the first in the tree to have the attribute "stop=1"
It could also be that page2 has a "stop=1" or any other page.

Any thoughts on what the xsl should look like?

Regards,
J. Diderik

The simple answer is like this:

<xsl:template match="page" priority="2">
  -- process the page
</xsl:template>

<xsl:template match="page[@stop=1] | page[following::page/@stop=1]"
       priority="3">
  // do nothing
</xsl:template>

However, this could be rather expensive, so whether it's viable depends on the size of your source. A faster solution, but one that's a bit harder to code, is to program the tree walking yourself (so a template for a page does apply-templates first to its first child, then to its following sibling, passing as a parameter a boolean that indicates whether stop=1 has yet been encountered.


Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thank you Michael,

I must first thank you for your book "XSLT Programmer's Reference and XPath 2.0 Programmer's Reference". It's been a great help in my struggles with xml.

I seem to be unable to get the simple solution to work, but will work on that. The courses will be quite large, with possibly over 100 pages, so walking the tree would be my best option, but I will have to dive into xsl all the way before I will get the full force of xsl to work :-)

Any steps in the right direction would be of great help.

YT,
J. Diderik

It occurs to me there's a way of doing it that's both simple and performant in XSLT 2.0.

First set a global variable to the first page with @stop=1

<xsl:variable name="stopper" select="(//page[@stop='1'])[1]"/>

Then in your template rules distinguish whether nodes are before or after the stopper:

<xsl:template match="page[. << $stopper]">

<xsl:template match="page[. is $stopper or . >> $stopper]"



Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Ah, Yes....all things fall into place :-)

I started to go down that road but didn't quite get it, your help is much appriciated!

Thank you very much.

Regards,
Jeroen Diderik